Background. The DSM-IV criteria for melancholia include the clinical feature ‘distinct quality’, defined as a mood state differing from that experienced in bereavement. Both propositions – its specificity to melancholia and its definition – remain problematical.

Methods. We examine both propositions by analysing an adjective checklist completed by melancholic and non-melancholic depressed subjects, as well as by a bereaved sample. The checklist was refined by a principal components analysis to four scales – one assessing a general ‘mood’ severity or dysphoric dimension, and the other three assessing dimensions of ‘fatigue’, ‘numbness’ and ‘guilt’.

Results. If the concept of ‘distinct quality’ has validity, we would require specificity of the refined qualitative constructs to melancholic depression. The ‘numbness’ component met that requirement, but only to a degree. While bereaved subjects did differ from those with melancholic depression on a number of our refined qualitative mood domains, such differences appeared more related to lower levels of depression in the bereaved sample.

Conclusions. We argue for deleting the ‘distinct quality’ criterion from diagnostic checklists of melancholia until its definition has been improved, its utility demonstrated and its specificity to any depressive subtype established as having clinical significance.